A progressively growing tumor is postulated to result after several steps of variation from normal (N). First, there is a normal (N) yields malignant (I) variation in which a cell arises sensitive to immune selection. Then, a second step variant, termed cancer (C) appears in this I population, which is resistant to some surveillance mechanisms; i.e., humoral or cell-mediated. The overall pathway of escape then is N yields I yields C where C is a series of variants resistant to different rejection mechanisms. Our objective is to isolate each of the single step variants and reconstruct the autochthonous pathway with them. The I variant will be isolated in tissue culture as a transformant. Variants can be isolated from I by imposing single-step selection in vitro using various effector mechanisms, IgG, NK or TK cell-mediated, IgM or IgG complement-mediated lysis, and in vivo using immune-depressed and normal mice. The properties of the I and C variants will be analyzed to determine: (1)\their cross-resistance patterns; (2) their ability to grow alone or in combinations as progressive tumors; (3) the number of different tumor-associated antigens expressed; (4) the number of gene level events required for the N yields I mutation and the I yields C mutation. A comparison of the resistance to immune effector attack and of the class of response induced will enable us to make general and predictive rules about tumor escape patterns. One particularly important comparison will be between the in vivo and the in vitro immunoselected derivatives.